Research On Location Privacy-Preserving Methods Against Background Knowledge Attack

With the widespread popularity of mobile communication devices and the advent of the5 G era,Location Based Service(LBS)as a new network service model enriches the growing material needs of users and is practically applied in military,medical,rescue,and other fields,which is of great significance for promoting the green development of the Internet.However,when users use location based services,they need to provide their location information and query attributes to the Location Service Provider(LSP).The contextual information attached to the location data and query attributes will expose the user’s privacy.Currently,user concerns about personal privacy are one of the important factors hindering the healthy development of the LBS industry.Therefore,researching location privacy protection methods under the premise of ensuring service quality is of great significance.Attackers can easily obtain background knowledge about users from various channels and use it for inference attacks based on map information,semantic information,and historical data,which can lead to sensitive information leakage.Existing location privacy protection methods based on location generalization are vulnerable to inference attacks by attackers with background knowledge,which poses a risk of location privacy leakage.Secondly,in a real road network environment,users have different privacy needs for different locations,and the privacy protection strength for different locations should meet actual needs,otherwise it can easily cause resource waste or inadequate protection.In response to these problems,this article combines historical query probability,location semantics,map information,and introduces differential privacy mechanism to carry out in depth research,and proposes a location privacy protection method to solve the above problems.The main research work of this article is as follows:(1)After detailed literature research,this article analyzed and summarized existing location privacy protection methods and pointed out the problems they face.To address these problems,this article proposes a new location privacy protection method that can protect users’ location privacy while ensuring service quality.(2)Analysis of the problems with existing location privacy protection methods for dummy location generation,and testing their security using multiple request attack algorithms.A multiple query attack algorithm was designed to test the security of existing dummy location generation methods for location privacy protection.An against background knowledge attack dummy location generation algorithm was proposed to effectively protect user location privacy.This method comprehensively considers background knowledge such as query probability,time distribution,location semantics,and physical dispersion to generate an effective set of dummy locations to resist probability distribution attacks,location semantics attacks,and location homogeneity attacks.When the user makes the first request,a special zero query location is transformed,and the dummy location set is constructed by selecting location units with similar query probabilities at the current request time using location entropy and time entropy.Then,location units that satisfy semantic differences are selected by adjusted cosine similarity.Finally,distance entropy is used to ensure a larger anonymity range between the selected location units,and the best dummy location set generated for the current request location is cached.Security analysis and simulation experiments show that the multiple query attack algorithm can identify the user’s true location with high probability,and the against background knowledge attack dummy location generation algorithm can effectively protect the user’s location privacy compared to existing dummy location generation methods.(3)An algorithm is proposed to address the issue of privacy leakage due to continuous queries,where an attacker can infer a user’s sensitive information based on their movement trajectory.The algorithm is designed to resist background knowledge attacks and provide personalized location privacy protection.First,a fuzzy mathematical model is used to address the issue of inconsistent sensitivity levels and thresholds for sensitive locations among different users.By combining a user’s historical query data,the algorithm determines the user’s sensitivity to different semantic location points.Next,a route with low privacy leakage and low travel cost is selected based on four factors: distance traveled,traffic congestion,road level,and total sensitive distance.Finally,the algorithm assigns a personalized privacy budget to the user based on their sensitivity and distance,and adds Laplace noise that conforms to differential privacy mechanisms to protect location privacy for the requested location point.Security analysis and experimental results show that the proposed algorithm can provide personalized location privacy protection for users while ensuring service quality.